Manuel M. M\"uller, Bj\"orn Bornkamp, Frank Bretz, Timothy I. Cannings, Wei Liu, Henry W. J. Reeve, Richard J. Samworth, Nikolaos Sfikas, Fang Wan, Konstantinos Sechidis
Data-driven controlled subgroup selection in clinical trials
https://arxiv.org/abs/2512.15676
Aditya Ghosh, Stefan Wager
Non-parametric Causal Inference in Dynamic Thresholding Designs
https://arxiv.org/abs/2512.15244
Jessica P. Kunke, Julian D. Olden, Tyler H. McCormick
Scalable Spatial Stream Network (S3N) Models
https://arxiv.org/abs/2512.12398
Antonio Olivas-Martinez, Peter B. Gilbert, Andrea Rotnitzky
Proximal Causal Inference for Modified Treatment Policies
https://arxiv.org/abs/2512.12038
Our framework unifies common measures like LOCO and Shapley values via a weighted sum of CATEs, and provides confidence bands and tests of zero importance, which is challenging because the null lies on the boundary of the parameter space.
We develop inference for model-agnostic, local variable importance for heterogeneous treatment effects, allowing importance to vary across individuals, while inference is global, testing whether a variable is important for any individual.
New paper: Inference on Local Variable Importance Measures for Heterogeneous Treatment Effects (with Peter B. Gilbert & @alexluedtke.bsky.social).
Preprint: arxiv.org/abs/2510.18843.
